Denon AVR-X3800H Review

[Narnian]

One aspect of the sub response problem in the modal region that's not widely discussed is the difference between the drywall construction that's typical in the U.S., and hard brick construction that's often found in Europe and elsewhere. This is mentioned briefly in the last sentence of the paragraph above that image in the documentation.

I am an odd one out as all of my walls are slightly rounded logs which should tend to disperse the reflection all over the place. Couple that with open rafter 8’-11’ ceiling rise from front to back I suspect that adds to a unique test result.

 

[peng]

I do plan on getting REW and the better mike to verify the Audyssey results at some point. I was expecting worse results to start and was surprised at the before measurements as to how relatively good they were from what I was expecting. The room is oddly shaped and arranged which helped break up the reflections. Looking for height speakers and will likely get REW then.

If you do that, and is willing to invest 2 to 3 hours you should be able to smooth the bass range more after running Audyssey as I have managed to do as shown below:

My DL PC standalone version vs XT32 vs nothing:
Note: 1/12 smoothing, frequency 25 to 300 Hz, front L+R, large, no sub, no manual adjustments, just as found:
That's the front L+R, set to "large".

After tweaking with the $20 App and the free Ratbuddyssey:
That's for the subwoofers 1 and 2 only:

All with 1/24 smoothing:

I don't know how much DL DLBC can improve, other than the impulse response, that I know DL definitely produced better results as measured, not sure if I could discern the difference.

 

[Narnian]

If you do that, and is willing to invest 2 to 3 hours you should be able to smooth the bass range more after running Audyssey as I have managed to do as shown below:

My DL PC standalone version vs XT32 vs nothing:
Note: 1/12 smoothing, frequency 25 to 300 Hz, front L+R, large, no sub, no manual adjustments, just as found:
That's the front L+R, set to "large".

After tweaking with the $20 App and the free Ratbuddyssey:
That's for the subwoofers 1 and 2 only:

All with 1/24 smoothing:

Thank you. Are these graphs generated by REW?

 

[peng]

Thank you. Are these graphs generated by REW?

Yes

 

[Sancus]

Room correction results for two SVS SB-1000. They are currently sitting under each L/R mains. I use the directional feature for music primarily. At some point I am going to add a third just because I have the older sub I haven’t sold yet. Very similar curves before and after likely due to placement.

This is interesting, first evidence I've seen that Audyssey does actually use the new hardware capability. In directional mode, it makes sense that it would naively EQ each sub separately. That is also what Dirac(w/o DLBC) does. It will produce worse results whether for music or movies, though. And there is very little music that actually has meaningfully independent bass sub-100hz.

I am skeptical because if "better results" means something like below as evidence, then I would like to see more evidence. If the goal is achieve what the graphs show, then I don't see anything remarkable about them versus what Audyssey+App can do.

What's remarkable about the graph is the reduction in seat variation. Audyssey does little to help seat to seat variation except by accident. If you want smooth response over multiple seats then all of those seats must have the same modes, and this is very unlikely. Even my small 3-person couch has a "bad seat" with different modes, and reclining/sitting up are both different too! Audyssey did not help my bad seat at all.

But even at the MLP, I've shown before that MSO produces better results with 2 subs. With my subs playing together, at the MLP, there is a big null from 40-60hz. Audyssey -- correctly -- leaves this null alone, because it has no way to fix it. MSO on the other eliminates it without any trouble. That's because MSO can use one sub with different room modes to compensate for the other. If Audyssey tries to boost BOTH subs, it will feed more energy into the null.

I am not doubting the MSO and DLBC's approach at all, just thought to ask the question, are those things overrated? I suppose the answer will most likely be, it depends..

Are my results a night and day difference at the MLP? No. Audible? Very much so, yes. Considering MSO is free(thanks @andyc56 !!!!) and I put the delay/gain/filters into my subs, for 0 cost and about 1 afternoon learning how MSO works I would say it's a very good result. Unfortunately I don't think I saved Audyssey-only vs MSO measurements for every seat, so you'll just have to take my word that my "bad seat" was also improved by MSO.

Whether or not it's overrated depends on your room, setup, and needs. If you have 4 subs and only care about the MLP, then yeah you probably don't need it. If you have a "real home theater" with multiple rows of seats and care about those seats? Yes I would say you 100% absolutely do need it.

Given the negligible cost, it's hard to see the downside other than time. But all of us are investing time posting on an audio forum, so we clearly don't value ours all that much As far as DLBC goes, it has a high cost and I would definitely NOT recommend paying for DLBC hardware/software for most people.

 

[peng]

This is interesting, first evidence I've seen that Audyssey does actually use the new hardware capability. In directional mode, it makes sense that it would naively EQ each sub separately. That is also what Dirac(w/o DLBC) does. It will produce worse results whether for music or movies, though. And there is very little music that actually has meaningfully independent bass sub-100hz.



What's remarkable about the graph is the reduction in seat variation. Audyssey does little to help seat to seat variation except by accident. If you want smooth response over multiple seats then all of those seats must have the same modes, and this is very unlikely. Even my small 3-person couch has a "bad seat" with different modes, and reclining/sitting up are both different too! Audyssey did not help my bad seat at all.

But even at the MLP, I've shown before that MSO produces better results with 2 subs. With my subs playing together, at the MLP, there is a big null from 40-60hz. Audyssey -- correctly -- leaves this null alone, because it has no way to fix it. MSO on the other eliminates it without any trouble. That's because MSO can use one sub with different room modes to compensate for the other. If Audyssey tries to boost BOTH subs, it will feed more energy into the null.

View attachment 249119​

Are my results a night and day difference at the MLP? No. Audible? Very much so, yes. Considering MSO is free(thanks @andyc56 !!!!) and I put the delay/gain/filters into my subs, for 0 cost and about 1 afternoon learning how MSO works I would say it's a very good result. Unfortunately I don't think I saved Audyssey-only vs MSO measurements for every seat, so you'll just have to take my word that my "bad seat" was also improved by MSO.

Whether or not it's overrated depends on your room, setup, and needs. If you have 4 subs and only care about the MLP, then yeah you probably don't need it. If you have a "real home theater" with multiple rows of seats and care about those seats? Yes I would say you 100% absolutely do need it.

Given the negligible cost, it's hard to see the downside other than time. But all of us are investing time posting on an audio forum, so we clearly don't value ours all that much As far as DLBC goes, it has a high cost and I would definitely NOT recommend paying for DLBC hardware/software for most people.

Thank you for the additional info. MSO is on my play list for sure. XT32 results do vary quite a bit among seats, but I happen to have just one seat lol so I am lucky.

Here's my 9 positions average that almost covered the single roll 3 seater couch:
The individual positions varied a lot, but the reference curve still produce better results than with Audyssey turned off.

 

[Zedly]

I also had excellent results using MSO to get ideal response from two subwoofers. The MLP is nearly perfect, and my three other seats are almost as good, except in the 75-100 Hz range where my room has issues.

 

[rvsixer]

To be useful, every FR measurement graph should include smoothing/software used. Otherwise anything can be made to look pretty.

 

[Chromatischism]

To be useful, every FR measurement graph should include smoothing/software used. Otherwise anything can be made to look pretty.

True, but at least for the REW screenshots above from @Zedly, the legend indicates no smoothing was used.

10-137 is sort of an arbitrary range, though. I use 10-200 or 10-300. Without as much stretching, the verticality of issues is more easy to see. I also think it makes sense to see the whole "bass" range, which I consider to go to around 250 give or take. If we all standardized our graphs would be easy to compare.

 

[rvsixer]

True, but at least for the REW screenshots above from @Zedly, the legend indicates no smoothing was used.

Point made. Not everyone here may immediately recognize the software packages the assortment of FR graphs being posted here come from, or especially know that a non-descript horizontal line in that unrecognized package's "legend" means no smoothing applied, etc.
Carry on...

 

[peng]

True, but at least for the REW screenshots above from @Zedly, the legend indicates no smoothing was used.

10-137 is sort of an arbitrary range, though. I use 10-200 or 10-300. Without as much stretching, the verticality of issues is more easy to see. I also think it makes sense to see the whole "bass" range, which I consider to go to around 250 give or take. If we all standardized our graphs would be easy to compare.

Yep, that's what I do too, if I don't specify smoothing it means no smoothing.

 

[peng]

Point made. Not everyone here may immediately recognize the software packages the assortment of FR graphs being posted here come from, or especially know that a non-descript horizontal line in that unrecognized package's "legend" means no smoothing applied, etc.
Carry on...

It is a very valid point. I would add one more, it is quite common that people would use the default Y axis scale, that also make the graphs look more impressive, though in this case the scales used are immediately recognizable. Zedly 55 to 95 scale is excellent! More people, especially reviewers, should do the same.

 

[pielover74]

For anyone thinking to pick up a better performing X3700H for less, you’re out of luck. Performance will be about the same because both units use the same DAC chip now.

I see nothing unexpected here. It basically performs only a bit worse than the DAC datasheet.

Would the x3700 or x2800 be a suitable replacement for my s760h? Sorry. I know this is an old comment

 

[delta76]

Would the x3700 or x2800 be a suitable replacement for my s760h? Sorry. I know this is an old comment

both are an upgrade. x is the premium line so they are a step up from s760h. if they're worth the extra money is up to you. why do you want to upgrade?

 

[pielover74]

both are an upgrade. x is the premium line so they are a step up from s760h. if they're worth the extra money is up to you. why do you want to upgrade?

Speaker bleed in surrounds when playing dd 2.0 content. I have two ls 50 metas in the front, r2c center svs sb 1000 and two klipsch rp 400m for surrounds. I live with the remote constantly in my hand because the volume sporadically changes. I’ve tried everything in the settings. Dynamic compression, etc and would hope a newer avr might fix these issues. Not sure if it’s not enough power being fed individually.

 

[delta76]

Not quite sure if I understand what "speaker bleed in surrounds" means, but yes that sounds like your s760h runs out of steam. if anything x3800h can be overkill (you probably won't need HDMI 2.1 or 4 subs out). you can go with x3600h or something similar (as long as it has Audyssey MultiQ XT32), and preouts so you can add an amp later if you want to.

 

[pielover74]

Not quite sure if I understand what "speaker bleed in surrounds" means, but yes that sounds like your s760h runs out of steam. if anything x3800h can be overkill (you probably won't need HDMI 2.1 or 4 subs out). you can go with x3600h or something similar (as long as it has Audyssey MultiQ XT32), and preouts so you can add an amp later if you want to.

Thanks for the replies! Speaker bleed is apparently a known issue with some Denon AVR in surrounds when playing DD 2.0 content unfortunately. I should’ve added, I keep the remote in hand because sometimes it gets louder than I’d expect and I can’t isolate if it’s because some speakers are getting more power than others or if it’s just the avr not being a great fit/or if it’s something I need to adjust

 

[delta76]

OK, it seems like a problem with upmixing 2.0 to 5.1 (I don't know if it's specific to Denon or not). did you run the calibration with Audyssey? they are supposed to fix that. (note that I mentioned XT32 because it's a big step up from XT, and that's a worthy upgrade)

 

[GalZohar]

Use the "direct" button multipld times until it selects "auto" mode and see if that helps, you might be running some undesireable surround upmixing and you should disable it before investigating further.

 

[ivo.f.doma]

OK, it seems like a problem with upmixing 2.0 to 5.1 (I don't know if it's specific to Denon or not). did you run the calibration with Audyssey? they are supposed to fix that. (note that I mentioned XT32 because it's a big step up from XT, and that's a worthy upgrade)

Yes, specific to Denon / Marantz.